Axiom Math
Startup building AI that writes and formally verifies math and code proofs in Lean.
On the record
Checkpoint: Whether Axiom names a paying enterprise customer with disclosed ARR in a regulated vertical
The call
Axiom has striking technical demos and a $1.6B valuation, but no named customer, no revenue, and a math-verification pitch that frontier labs could absorb into their own products.
Axiom Math builds software that checks whether artificial intelligence has solved math problems correctly, and it has published strong technical results to prove the technology works. The company's bet is that banks, healthcare companies, and government agencies will eventually be required by law to independently verify any math or code that AI generates for them. However, the bigger AI labs like OpenAI and DeepMind are already building math-checking directly into their own AI systems, and Axiom has no paying customers yet to show the regulatory requirement is real. If these frontier labs keep improving their math reasoning on their own, companies may never feel the need to buy Axiom's separate verification tool at all.
Why
01
AI pressure: OpenAI, DeepMind, Anthropic. Frontier labs already post gold-level math reasoning; if they bundle native Lean-checked output, Axiom's whole pitch shrinks. Could AI do the core job? Yes, likely.
02
Bigger players: OpenAI, DeepMind, Anthropic. Same labs could fold formal verification into their own coding tools, skipping a dedicated vendor like Axiom entirely. Could a giant copy it? Yes, easily.
03
What protects it. Deep Lean-specific tooling (AXLE) and a formal verification pipeline that generalist model labs have not built into their products yet.
What would change this call
Axiom names an enterprise customer with disclosed ARR in finance, defense, or chip design within 12 months. Would show the formal-verification moat is converting into paying regulated customers.
OpenAI, DeepMind, or Anthropic ships a native Lean-checked or verified-output feature bundled into their model. Would show frontier labs substituting for Axiom's core value proposition.
Harmonic announces a named customer win or a benchmark result claimed to beat Axiom's Lean scores. Would show the nearest identical peer pulling ahead in the same target market.
OpenAI, DeepMind, or Anthropic ships a native Lean-checked or verified-output feature bundled into their model. Would show frontier labs substituting for Axiom's core value proposition.
Harmonic announces a named customer win or a benchmark result claimed to beat Axiom's Lean scores. Would show the nearest identical peer pulling ahead in the same target market.
Key numbers
- Series A: $200M at $1.6B valuation
- Claims 98.93% on Lean benchmark vs 11-12% rivals
- $0 disclosed revenue despite $264M raised
Named risks
- No disclosed revenue or named enterprise customers despite $1.6B valuation as of mid 2026.
- Core value prop of needing formal verification is undercut as frontier labs close the math-accuracy gap themselves.
- Direct, equally-funded rival Harmonic pursues the identical Lean-verification strategy with no clear differentiation yet.
- Funding and valuation figures show minor inconsistencies across trackers, with one figure originally single-sourced to a YouTube video.
Rivals, as read in July 2026
| Rival | Direction | Evidence |
|---|---|---|
| Harmonic | contesting | Both raced to same $1.4-1.6B valuation range in late 2025/early 2026 for formalization vendor slot |
| OpenAI | gaining on | Reached gold-medal IMO level in July 2025 via general reasoning, no formal tooling needed |
| Google DeepMind | gaining on | AlphaProof already uses Lean and hit gold-level IMO reasoning by 2025 |
| Anthropic | gaining on | Fable 5 model reportedly produced a Lean-checked proof in July 2026 |
Questions people ask
- Is Axiom Math profitable in 2026?
- No. Axiom Math has disclosed zero revenue as of February 2026 despite raising a $200M Series A at a $1.6B valuation in March 2026 and reaching roughly $264M in total funding by July 2026, per Sacra and TrueUp trackers.
- Who are Axiom Math's biggest competitors?
- Harmonic, a nearly identical Lean-based formal-verification startup valued at $1.45B in November 2025, is Axiom's closest rival. Frontier labs OpenAI, Google DeepMind, and Anthropic also ship competing math-reasoning capability without dedicated formal tools.
- Could AI make Axiom Math's product unnecessary?
- Possibly. OpenAI and DeepMind both reached gold-medal IMO level in July 2025 using general reasoning, not formal verification, and Anthropic reportedly produced a Lean-checked proof in July 2026, chipping at Axiom's core need-for-verification pitch.
- How much money has Axiom Math raised?
- Axiom Math raised a $64M seed round in October 2025 at a roughly $300M valuation, then a $200M Series A in March 2026 at a $1.6B valuation led by Menlo Ventures, with total funding reported at $264M as of July 2026.
- Does Axiom Math have any paying customers?
- Fortune reported in June 2026 that Axiom Math has customers in finance, cybersecurity, and software or hardware development, but named no specific companies or dollar figures, and no disclosed revenue has been reported elsewhere.
- What has Axiom Math actually built?
- Axiom Math built AxiomProver, a Lean-based formal reasoning engine that scored a perfect 120/120 on the December 2025 Putnam competition and produced proofs accepted into five mathematical journals as of June 2026.
Timeline
- 2026-07-28Anthropic's Fable 5 model reportedly produces a Lean-checked proof breakthrough, per Forbes
- 2026-06-01Fortune reports ~40 employees and unnamed customers in finance, cybersecurity, software/hardware
- 2026-05-13Axiom claims 98.93% on a Lean verification benchmark versus 11-12% for alternatives (vendor-claimed)
- 2026-03-12Series A of $200M closes at $1.6B valuation led by Menlo Ventures
- 2025-12-31AxiomProver scores a perfect 120/120 on the Putnam Competition
- 2025-11-30Rival Harmonic raises $120M Series C at $1.45B valuation
- 2025-10-02Seed round of $64M closes at ~$300M valuation led by B Capital
- 2025-07-31OpenAI and DeepMind both reach gold-medal IMO level without dedicated formal tooling
Revision history
29 July 2026
First assessed. Losing its protection. The text above is fixed as of this date and is not edited afterwards.
In the same band
Sources
- Verifiable AI startup Axiom raises $200M to prove AI-generated code is safe to use (2026-03-12)
- Start-ups are racing to revolutionise mathematics with AI (2026-06-03)
- AGI Needs Formal Reasoning. Carina Hong is Building it at Axiom. (2026-05-13)
- AI startup Axiom gets $64M to develop new knowledge with advanced mathematics (2025-10-02)
- Company Profile
- Exclusive: Economists have been teaching an unproven proof for 50 years. AI just solved it (2026-06-01)
- Axiom Math valuation, funding & news
- Mathematicians say AI cracked research proofs. The best tools aren't public. (2026-04-14)
- AI achieves silver-medal standard solving IMO problems, Google DeepMind (2026-03-03)
- AI startup tackles bottleneck where people spend more time checking AI content than creating it (2025-07-29)
- Sam Altman Says We're In The Singularity. What Does He Actually Mean? (2026-07-28)
- Toward Mathematical Superintelligence: Why We Invested in Axiom (2026-01-21)